A new screening test app could help advance the early detection of Parkinson’s disease and severe COVID-19, improving the management of these diseases.
Developed by a research team of engineers and neurologists led by RMIT University in Melbourne, the test can produce accurate results using only people’s voice recordings.
Millions of people around the world have Parkinson’s disease, a degenerative brain disease that can be difficult to diagnose because symptoms vary from person to person. Common symptoms include slow movements, tremors, stiffness, and imbalance.
Currently, Parkinson’s disease is diagnosed through an evaluation by a neurologist which can take up to 90 minutes.
Powered by artificial intelligence, the smartphone app records a person’s voice and takes just 10 seconds to reveal if they may have Parkinson’s disease and should be referred to a neurologist.
Lead researcher Professor Dinesh Kumar, from RMIT’s School of Engineering, said the easy-to-use screening test made it ideal for use in a national screening programme.
He said the team had developed a similar test for people with COVID-19 to reveal if they needed clinical attention, including hospitalization.
“Early detection, diagnosis and treatment could help manage these diseases, so making screening faster and more accessible is critical,” Kumar said.
“This research will enable a non-contact, easy-to-use, low-cost test that can be performed routinely anywhere in the world, where clinicians can monitor their patients remotely.
“It could also promote a community-wide testing program, reaching people who otherwise wouldn’t seek treatment until it’s too late.”
Research results are published in IEEE Journal of Translational Engineering in Health and Medicine, IEEE Access and Computers in biology and medicine.
How the technology works
The voice of people with Parkinson’s disease changes due to a combination of three symptoms: rigidity, tremors and sluggishness (known as bradykinesia). Expert clinicians can identify these symptoms, but this assessment can be difficult due to the large natural differences in people’s voices.
Kumar said previous attempts to develop a computerized voice assessment to detect Parkinson’s disease had been inaccurate because of these significant differences in people’s voices.
“As part of our research, we used voice recordings of people with Parkinson’s disease and a control group of so-called healthy people uttering three sounds – A, O and M – which are similar to the singing of Hindu meditation,” Kumar said.
“These sounds allow for more accurate disease detection.”
In patients with symptoms of lung disease due to COVID-19, there is also voice change due to lung infection, Kumar said.
“Again, due to the large differences in people’s voices, lung disease is difficult to recognize in its early stages,” he said.
“We overcame this limitation with the choice of these same three sounds and the AI analysis method we developed.”
Before being used, the system is trained to identify the disease. Once trained, it performs instant voice analysis.
The software then compares the results with existing samples of voices from people with Parkinson’s disease and those without.
Co-researcher Dr Quoc Cuong Ngo from RMIT’s School of Engineering said the new technology was faster and better than any similar AI-based approach.
“Our screening test app can measure, with great precision, how the voice of someone with Parkinson’s disease or someone at high risk of hospitalization due to COVID-19 is different from healthy people,” he said.
The team wants to do a larger observational study to detect the progression of Parkinson’s disease and lung disease.
“We also want to test the effectiveness of this technology for other diseases, such as other neurological conditions and sleep disorders,” Kumar said.
“We are looking for a commercial partner and a clinical partner ahead of a clinical trial planned for next year.”
Pah ND, Indrawati V, Kumar DK. Supported phoneme speech characteristics as a COVID-19 biomarker. IEEE Journal of Translational Engineering in Health and Medicine. 2022;10:1-9. doi:10.1109/JTEHM.2022.3208057
This article was republished from the following documents. Note: Material may have been edited for length and content. For more information, please contact the quoted source.